Technologies for classification and clustering Clause Samples

Technologies for classification and clustering. Several frameworks and libraries are already available for classification and clustering. In the following, we introduce the most famous ones: 16 ▇▇▇▇://▇▇▇▇▇▇.▇▇▇▇▇▇.▇▇▇/ 18 ▇▇▇▇▇://▇▇▇.▇▇▇▇▇▇.▇▇▇/ 19 ▇▇▇▇▇://▇▇▇▇▇▇.▇▇▇/lyst/lightfm 20 ▇▇▇▇://▇▇▇.▇▇▇▇▇▇▇▇▇▇▇.▇▇▇ 21 ▇▇▇▇▇://▇▇▇▇▇▇.▇▇▇/RankSys/RankSys ● Apache Mahout22 (Apache License v2) includes a Java implementation for popular machine learning algorithms for classification, clustering and topic modelling. The library is highly scalable, and it is characterized by having a good documentation, high performance, and maintainability. ● Scikit-learn23 (BSD license) offers a Python implementation of the most popular machine learning algorithms for classification and clustering. The library is characterized by an excellent documentation, high performance, ease of use and maintainability. The drawbacks are scalability and heavy multiprocessing instead of lightweight multithreading (due to Python GIL limitations - see ▇▇▇▇▇://▇▇▇▇.▇▇▇▇▇▇.▇▇▇/moin/GlobalInterpreterLock). ● Weka24 (GNU General Public License), developed in Java, supports diverse machine learning approaches for classification, and clustering and attribute selection. In addition to the graphical user interface provided, it can be accessed via a Java API.